The long-term goal of this proposal is to define the role of DNMT3A mutations for the pathogenesis of acute myeloid leukemia (AML), so that rational targeted therapies can be designed to counteract these mutations. DNMT3A (which encodes a de novo DNA methyltransferase) is among the most commonly mutated genes in AML (22% of all patients, and 34% of intermediate risk). Most mutations are missense, and the most common of these (60% of all cases) occurs at amino acid R882. However, many AML genomes have loss-of-function mutations in DNMT3A that are reminiscent of inactivating mutations in classical tumor suppressors. All DNMT3A mutations are associated with adverse overall survival. It is critical to understand whether the common DMNT3A mutations (especially at R882) are acting as gain-of-function or loss-of-function alleles, since this information will influence approaches to targeted therapies. We will assess the biochemical consequences of DNMT3A mutations in Aim 1, and the in vivo consequences in Aims 2 and 3, as follows: Specific Aim 1: We will define how AML-associated mutations alter DNMT3A functions in vitro. Recombinant DNMT3A protein with R882 (and other) mutations will be produced in E. coli, and tested for alterations in cytosine methylase activity and specificity, and binding to other cellular proteins. We will determine whether specific mutations cause loss-of-function or gain-of-function properties (and/or have dominant negative activities), and use this information to select additional mutations for study in vivo. Specific Aim 2: We will determine whether the DNMT3A R882H mutation can contribute to AML pathogenesis in vivo. Using retroviral vectors with fluorescent tags, we will express WT and mutant R882H DNMT3A cDNAs in murine bone marrow cells. Transduced cells will be evaluated in colony assays and mice to assess the leukemogenic potential of DNMT3A mutations alone or in combination with other mutations that commonly occur with DNMT3A mutations (e.g. FLT3 ITD, IDH1 R132H, NPMc). Using homologous recombination in ES cells, we will create mice with the R882H mutation at the identical position in the mouse Dnmt3a gene (R878H), and assess development, hematopoiesis, and cancer susceptibility. Specific Aim 3: We will assess the effect of Dnmt3a haploinsufficiency on AML pathogenesis. Mice haploinsufficient for Dnmt3a do not have a measurable hematopoietic phenotype. We will perform tumor watches using mice that are haploinsufficient for Dnmt3a, and we will compare the leukemogenicity of FLT3 ITD, NPMc, and IDH1/2 mutations using WT vs. Dnmt3a haploinsufficient bone marrow cells.